Performance helpers

Numba-accelerated retarded integrator utilities.

This module is a faithful, structured transcription of legacy/numba_optimized_integrator.py. The goal is to expose the validated optimised routines in a predictable API while leaving the original legacy file untouched for regression comparison.

class core.performance.OptimisationOptions(use_numba: bool = True, run_benchmark: bool = False, self_consistency: SelfConsistencyConfig | None = None)

Bases: object

Control flags for run_optimised_integrator().

use_numba: bool = True

run_benchmark: bool = False

self_consistency: SelfConsistencyConfig | None = None

__init__(use_numba: bool = True, run_benchmark: bool = False, self_consistency: SelfConsistencyConfig | None = None) → None

core.performance.dict_to_arrays(particle_dict: Dict[str, ndarray]) → Tuple[Dict[str, ndarray], int]

core.performance.arrays_to_dict(arrays: Dict[str, ndarray]) → Dict[str, ndarray]

core.performance.eqsofmotion_retarded_numba(h: float, trajectory, trajectory_ext, index_traj: int, aperture_radius: float, sim_type: SimulationType) → Dict[str, ndarray]

core.performance.retarded_integrator_numba(steps: int, h_step: float, wall_z: float, aperture_radius: float, sim_type: SimulationType, init_rider: Dict[str, ndarray], init_driver: Dict[str, ndarray] | None, mean: float, cav_spacing: float, z_cutoff: float, self_consistency: SelfConsistencyConfig | None = None) → Tuple[Tuple[Dict[str, ndarray], ...], Tuple[Dict[str, ndarray], ...]]

core.performance.run_optimised_integrator(config: IntegratorConfig, init_rider: Dict[str, ndarray], init_driver: Dict[str, ndarray] | None, options: OptimisationOptions | None = None) → Tuple[Tuple[Dict[str, ndarray], ...], Tuple[Dict[str, ndarray], ...]]